RESEARCH ARTICLE Open Access

Raloxifene inhibits tumor growth and lymphnode metastasis in a xenograft model ofmetastatic mammary cancerMasa-Aki Shibata1*, Junji Morimoto2, Eiko Shibata1,6, Hitomi Kurose1, Kanako Akamatsu3, Zhong-Lian Li1,Moriaki Kusakabe4, Masahide Ohmichi5, Yoshinori Otsuki1*

Abstract

Background: The effects of raloxifene, a novel selective estrogen receptor modulator, were studied in a mousemetastatic mammary cancer model expressing cytoplasmic ERa.Methods: Mammary tumors, induced by inoculation of syngeneic BALB/c mice with BJMC3879luc2 cells, weresubsequently treated with raloxifene at 0, 18 and 27 mg/kg/day using mini-osmotic pumps.

Results: In vitro study demonstrated that the ERa in BJMC3879luc2 cells was smaller (between 50 and 64 kDa)than the normal-sized ERa (66 kDa) and showed cytoplasmic localization. A statistically significant but weakestradiol response was observed in this cell line. When BJMC3879luc2 tumors were implanted into mice, the ERamRNA levels were significantly higher in females than in males. In vitro studies showed that raloxifene inducedmitochondria-mediated apoptosis and cell-cycle arrest in the G1-phase and a decrease in the cell population in theS-phase. In animal experiments, tumor volumes were significantly suppressed in the raloxifene-treated groups. Themultiplicity of lymph node metastasis was significantly decreased in the 27 mg/kg group. Levels of apoptosis weresignificantly increased in the raloxifene-treated groups, whereas the levels of DNA synthesis were significantlydecreased in these groups. No differences in microvessel density in tumors were observed between the controland raloxifene-treated groups. The numbers of dilated lymphatic vessels containing intraluminal tumor cells weresignificantly reduced in mammary tumors in the raloxifene-treated groups. The levels of ERa mRNA in mammarytumors tended to be decreased in the raloxifene-treated groups.

Conclusion: These results suggest that the antimetastatic activity of raloxifene in mammary cancer expressingcytoplasmic ERa may be a crucial finding with clinical applications and that raloxifene may be useful as anadjuvant therapy and for the chemoprevention of breast cancer development.

BackgroundThe selective estrogen receptor modulators (SERMs)exhibit specific estrogen-receptor (ER) agonistic andantagonistic activity by binding to ERa and/or b. Of theSERMs, tamoxifen and raloxifene differ from estrogensin that they exert both agonistic and antagonistic prop-erties. Tamoxifen acts as an antagonist in the breast andan agonist in the bone and uterus. Therefore, tamoxifenis used clinically as a therapeutic agent to treat ER-

positive breast cancer. Although tamoxifen prevents ER-positive breast cancers [1], it increases the incidence ofendometrial cancer [2,3]. Raloxifene has antiestrogeniceffects on the breast and bone, but it does not have anestrogenic effect on the uterus. These SERMs have dif-ferent biological actions because raloxifene recruits aco-repressor in endometrial carcinoma cells, whereastamoxifen induces a co-activator [4]. In fact, raloxifeneinhibits carcinogen-induced mammary carcinoma [5-7]and colon carcinoma [8] in animal models.The Study of Tamoxifen and Raloxifene (STAR) trial

has shown that raloxifene is as effective as tamoxifen inreducing the risk of invasive breast cancer, and there

* Correspondence: shibatam@art.osaka-med.ac.jp; an1001@art.osaka-med.ac.jp1Department of Anatomy and Cell Biology, Division of Life Sciences, OsakaMedical College, 2-7 Daigaku-machi, Takatsuki, Osaka 569-8686, JapanFull list of author information is available at the end of the article

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

© 2010 Shibata et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.

were less cases of endometrial cancer with raloxifenethan with tamoxifen [9]. Results of other clinical trialsof raloxifene, such as the Multiple Outcomes of Raloxi-fene Evaluation (MORE) [10], Continuing OutcomesRelevant to Evista (CORE) [11] and the Raloxifene Usefor The Heart (RUTH) [12] trials, showed that raloxi-fene reduces the risk of invasive ER-positive breast can-cer in postmenopausal women. As compared withtamoxifen, raloxifene appears to have fewer serious sideeffects, including endometrial cancer, venous thrombosisand cataracts, without compromising the breast cancerchemoprevention strategy [13].Breast cancer is the most common malignancy in

women worldwide and is one of the most lethal carcino-mas. In Japan, the incidence of breast cancer is continu-ously increasing and the disease now ranks fifth as acause of female mortality; the number of breast cancerdeaths in Japan increased 2.6-fold between 1975 and1998 [14]. The lethality of breast cancer is largely due tometastasis; the most common sites are lung, lymphnodes, liver, and bone. Effective and less toxic chemo-preventive agents are needed to delay the progression ofbreast cancer and prolong life.Here, we investigated the chemopreventive ability of

raloxifene, especially its antimetastatic ability, in amouse metastatic mammary cancer model expressingcytoplasmic ERa. This mammary cancer model has ap53 mutation that shows a metastatic spectrum similarto that seen in human breast cancers [15-17]. In addi-tion, we studied the apoptosis pathway, DNA synthesis,and cell cycle in metastatic mouse mammary carcinomacells treated with raloxifene in vitro.

MethodsExperimental regimenRaloxifene hydrochloride was purchased from Sigma Co.(St. Louis, MO, USA). For in vitro use, raloxifene wasdissolved in dimethylsulfoxide (DMSO), and aliquots of20 mM stock solution were stored at -20°C.

Cell line and animalsThe BJMC3879luc2 mammary carcinoma cell line [18]was generated by stable transfection of luc2 (animproved firefly luciferase gene) into parent cell lineBJMC3879. The mammary tumors arising fromBJMC3879 cell implantation had a high propensity formetastasis into the lymph nodes and lungs [15-17], atrait retained through culture. BJMC3879luc2 cells weremaintained in RPMI 1640 medium containing 10% fetalbovine serum with streptomycin/penicillin in an incuba-tor under 5% CO2.Thirty female 6-week-old BALB/c mice were used in

this study (Japan SLC, Hamamatsu, Japan). The animalswere housed five per plastic cage on wood chip bedding

with free access to water and food under controlledtemperature (21 ± 2°C), humidity (50 ± 10%), and light-ing (12-12 h light-dark cycle). All animals were held fora 1-week acclimatization period before study com-mencement. Mice were treated in accordance with theprocedures outlined in the Guide for the Care and Useof Laboratory Animals in Osaka Medical College, theJapanese Government Animal Protection and Manage-ment Law (No. 105) and the Japanese GovernmentNotification on Feeding and Safekeeping of Animals(No. 6).

Estrogen receptor expressionImmunofluorescence stainingBJMC3879luc2 cells were grown in 2-well chamberslides and fixed in 4% formaldehyde solution in phos-phate buffer. Immunofluorescence staining was per-formed with anti-ERa rabbit polyclonal antibody (cloneMC-20; Santa Cruz Biotechnology, Santa Cruz, CA,USA).Western blottingTotal protein was extracted from whole cell lysates ofBJMC3879luc2 cells. Total protein (40 μg) was electro-phoretically separated in 14% Tris-glycine gels under redu-cing conditions and transferred to nitrocellulosemembranes. The membrane was incubated with anti-ERa(Santa Cruz Biotechnology) or anti-ERb (Affinity Biorea-gents, Golden, CO, USA) rabbit polyclonal antibodies, fol-lowed by secondary antibodies conjugated to HRP. Then,the bound antibody was visualized with enhanced chemi-luminescence reagent (Perkin Elmer Life Sciences Inc.,Boston, MA, USA). Blots were visualized using aLAS-3000 image analyzer (Fujifilm, Co., Tokyo, Japan).Anti-Bid and anti-actin goat polyclonal antibodies (SantaCruz Biotechnology) were used as primary antibodies.ERa expression of BJMC3879luc2-implanted tumors infemales and malesBJMC3879luc2 cells (5 × 106 cells/0.3 ml in PBS) wereinoculated subcutaneously into the right inguinal mam-mary fat pad of 10 BALB/c mice (5 females and 5males). From two to four weeks after the inoculation,the mammary tumors were measured with digital cali-pers. The tumor volumes were calculated with the fol-lowing formula: maximum diameter × (minimumdiameter)2 × 0.4 [19]. Four weeks after inoculation,mammary tumors were immediately excised under iso-flurane anesthesia. Total RNA was extracted, and tran-scriptional levels of ERa were measured in mammarytumors using real-time reverse transcriptase (RT)-PCR(see “ERa expression in mammary tumors” for details).

Cell viabilityBJMC3879luc2 cells were grown in RPMI-1640 mediumsupplemented with 10% (v/v) heat-inactivated fetal

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 2 of 14

bovine serum and 2 mM L-glutamine under an atmo-sphere of 95% air and 5% CO2 at 37°C. BJMC3879luc2cells were plated onto 96-well plates (1 × 104 cells/well)1 day before raloxifene treatment. They were subse-quently incubated for 24 h with culture medium con-taining vehicle (DMSO) alone or with mediumcontaining raloxifene at different concentrations up to80 μM. Cell viability was determined using a CellTiter-Blue Cell Viability Assay (Promega Co., Madison, WI,USA). In addition, to examine the response to estrogenin BJMC3879luc2 cells, one week before the experiment,the medium was changed to a phenol red-free form ofRPMI-1640 containing charcoal-stripped fetal bovineserum (steroid-free medium). The cells were seeded at 5× 104 cells/well in the culture medium (steroid-free andphenol red-free). After overnight culture, the cells wereexposed to 17-b estradiol (E2) at final concentrations of10-12 to 10-4 M (1 pmole to 100 μmole) for 24 h, andthen cell viability was determined as described above.

TUNEL assay, caspase activity and DNA synthesisBJMC3879luc2 cells were grown in 2-well chamber slidesand treated with 20 μM raloxifene for 48 h. Then, the cellswere fixed in 4% formaldehyde solution in phosphate buf-fer, and terminal deoxynucleotidyl transferase-mediateddUTP-FITC nick end-labeling (TUNEL) staining was per-formed according to the manufacturer’s protocol (WakoPure Chemical Industries, Osaka, Japan).BJMC3879luc2 cells were plated onto 96-well plates (1

× 104 cells/well) 1 day before raloxifene treatment. Cellswere treated with 20 μM raloxifene or vehicle alone for48 and 72 h, and then cell viability was measured usinga CellTiter-Blue Cell Viability Assay (Promega). Theactivities of caspase-8, caspase-9 and caspase-3 weremeasured using a luminescent assay kit (Promega). Cas-pase activity was measured in terms of the luminescentsignal produced by caspase cleavage of the correspond-ing substrate using a Luminoskan Ascent device(Thermo Electron Co., Helsinki, Finland). Caspase activ-ity levels were corrected by the corresponding cellviabilities. In addition, cells from the cultures were incu-bated for 1 h in medium containing 50 μM 5-bromo-2’-deoxyuridine (BrdU), and DNA synthesis of the cellswas measured by BrdU incorporation (Cell Proliferationon ELISA, BrdU Chemiluminescence; Roche Diagnos-tics, GmbH, Mannheim, Germany). Data were also cor-rected by the corresponding cell viabilities.

Release of cytochrome cAfter incubation in culture medium with or without 20μM raloxifene for 48 h, both floating and attached cellswere harvested, rinsed once in PBS, re-suspended in celllysis buffer, incubated for 1 h at room temperature, andcentrifuged at 1000 × g for 15 min. The resultant

supernatant was diluted at least 5-fold. Supernatantscontaining the cytosolic fraction were collected sepa-rately, and the protein concentrations were determined.To determine the cytochrome c release into the cytosol,cytochrome c was measured using a cytochrome c kit(R&D Systems, Inc, Minneapolis, MN, USA).

Caspase inhibitor experimentCells were treated with 10 μM and 100 μM of the fol-lowing caspase inhibitors for 48 h: z-VAD-fmk againstbroad-spectrum caspases, Ac-DNLD-CHO against cas-pase-3, z-IETD-fmk against caspase-8 and z-LEHD-fmkagainst caspase-9. The caspase inhibitors, with theexception of caspase-3 inhibitor (Peptide Institute, Inc.,Osaka, Japan), were purchased from MBL Inc. (Nagoya,Japan). Although DEVD has been generally used as acaspase-3 inhibitor, this sequence has been reported tobe non-specific to caspase-3; therefore, Ac-DNLD-CHOwas used in the present experiment [20,21]. Two hoursafter treatment with caspase inhibitors, cells wereexposed to 20 μM raloxifene. Cell viability was mea-sured using a fluorescent assay kit (CellTiter-Blue CellViability Assay, Promega), and then the activities of cas-pase-3, caspase-8 and caspase-9 were measured using aluminescent assay kit (Promega). The caspase activitydata was then adjusted to account for the correspondingcell viability as previously reported [22].

Cell-cycle distributionFlow cytometric analysis was conducted on trypsinizedBJMC3879luc2 cell suspensions that were harvestedafter a 48-h treatment with 20 μM raloxifene and fixedin cold 70% ethanol. The cells were stained with a 50μg/ml propidium iodide solution containing 100 μg/mlRNase A for 30 min at 37°C and then placed on ice justprior to flow cytometric analysis (EPICS Elite ESP; Coul-ter Co., Miami, FL, USA). The percentage of cells ineach phase of the cell cycle was determined using aMulticycle Cell Cycle Analysis program (Coulter Co.).

In vivo study of raloxifene in a metastatic mammarycancer modelTwo dosages of raloxifene for mice (27 mg/kg and 18mg/kg) were selected based on the results of other stu-dies [23]. Raloxifene was continuously administered viasubcutaneously implanted mini-osmotic pumps (Alzetmodel 2002, Durect Co., Cupertino, CA, USA) that werecalibrated to release 0.5 μl of solution per hour. Raloxi-fene solutions (47.5 mg/ml and 31.7 mg/ml) in DMSOand 100% ethanol (1:3, v/v) were prepared. Since thepumps were calibrated to release for 14 days, they werereplaced every other week.BJMC3879luc2 cells (5 × 106 cells/0.3 ml in PBS) were

subcutaneously inoculated into the right inguinal

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 3 of 14

mammary fat pad of 30 female BALB/c mice. Two weekslater, when tumors had reached approximately 0.6 cm indiameter, mini-osmotic pumps were used to administer 0,18 or 27 mg/kg raloxifene for 6 weeks. Individual bodyweights were recorded weekly. Each mammary tumor wasalso measured weekly using digital calipers, and tumorvolumes were calculated according to the formula of maxi-mum diameter × (minimum diameter)2 × 0.4 [19]. All ani-mals received 50 mg/kg BrdU (Sigma Co.) i.p. at 1 h priorto sacrifice. All surviving mice were euthanized with iso-flurane anesthesia at week 6.

Bioluminescence imaging in vivoAt week 6, five mice in each group were anesthetized byisoflurane inhalation with an SBH Scientific anesthesiasystem (SBH Designs Inc., Ontario, Canada). Eachanesthetized mouse received an intraperitoneal injectionof 3 mg of D-luciferin potassium salts (Wako Pure Che-mical Industries). Bioluminescence imaging with aPhoton Imager (Biospace Lab, Paris, France) was per-formed. The bioluminescent signals received during the6-min acquisition time were quantified using Photovi-sion software (Biospace Lab).

Histopathological analysesAt necropsy, tumors and lymph nodes were removed,fixed in 10% formaldehyde solution in phosphate bufferand processed through to paraffin embedding. Thelymph nodes from the axillary and femoral regions wereroutinely removed, along with lymph nodes thatappeared abnormal. In several cases, the uterus was alsoexcised and preserved in fixative solution. Lungs wereinflated with formaldehyde solution prior to excisionand immersion in fixative; the individual lobes were sub-sequently removed from the bronchial tree and exam-ined for metastatic foci and similarly processed throughto paraffin embedding. All paraffin-embedded tissueswere cut into 4-μm-thick sections. Sequential sectionswere stained with hematoxylin and eosin for histopatho-logical examination or remained unstained for immuno-histochemical analysis.

p53 immunohistochemistryThe labeled streptavidin-biotin (LSAB) method (Dako,Glostrup, Denmark) was used for p53 immunohisto-chemistry. Unstained sections were immersed in distilledwater and heated for antigen retrieval prior to incuba-tion with a p53 mouse monoclonal antibody (ClonePab240, Santa Cruz Biotechnology) that reacts to themutant protein in fixed specimens.

Apoptosis and caspase in mammary tumorsFor the quantitative analyses of cell death, sectionsfrom paraffin-embedded tumors were assayed using

the TUNEL method in conjunction with an apoptosisin situ detection kit (Wako Pure Chemical Industries)with minor modifications to the manufacturer’s proto-col. TUNEL-positive cells (mainly regarded as apopto-tic cells) were counted in viable regions peripheral toareas of necrosis in tumor sections. The slides werescanned at low-power (×100) magnification to identifythose areas having the highest number of TUNEL-positive cells. Four areas neighboring the highest areaof TUNEL-positive cells were then selected andcounted at higher (×200-400) magnification. The num-bers of TUNEL-positive cells were expressed as num-bers per cm2.Active caspase expression of the mammary tumor tis-

sues was immunohistochemically detected using anti-cleaved caspase-3 and cleaved caspase-9 rabbit polyclonalantibodies (Cell Signaling Technology, Danvers, MA,USA). Immunohistochemistry was conducted using theLSAB method, and CSA II amplification (Dako) wasadditionally applied to detect cleaved caspase-9.

DNA synthesis in mammary tumorsThe tumors from five animals from each treatmentgroup were subsequently evaluated for DNA synthesisrates as inferred by BrdU incorporation. DNA wasdenatured in situ by incubating unstained paraffin-embedded tissue sections in 4 N HCl solution for 20min at 37°C. The incorporated BrdU was visualizedafter exposure to an anti-BrdU mouse monoclonalantibody (Clone Bu20a, Dako). The numbers of BrdU-positive S-phase cells per 250 mm2 were counted infour random high-power (×400) fields of viable tissue,and the BrdU labeling indices were expressed as num-bers per cm2.

Lymphatic and blood microvascular densities inmammary tumorsTo quantitatively assess lymphatic and blood microves-sel density in the primary mammary carcinomas, immu-nohistochemistry based on the LSAB method (Dako)was performed. A hamster anti-podoplanin monoclonalantibody (AngioBio Co., Del Mar, CA, USA) against alymphatic endothelium marker and a rabbit polyclonalantibody against CD31 (Lab Vision Co., Fremont, CA,USA), a specific marker for blood vessel endothelium,were used. The number of podoplanin-positive lympha-tic vessels containing intraluminal tumor cells was alsocounted. In addition, the number of CD31-positiveblood microvessels was counted as previously described[24]. Briefly, the slides were scanned at low-power(×100) magnification to identify those areas having thehighest number of vessels. The five areas of highestmicrovascular density were then selected and counted athigher (×200-400) magnification.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 4 of 14

ERa expression in mammary tumorsImmunohistochemical staining for ERa (anti-ERa rabbitpolyclonal antibody, Santa Cruz Biotechnology) was per-formed using the LSAB method in combination with aCSA II amplification kit (Dako). In addition, the levelsof ERa mRNA in mammary tumor tissues were alsomeasured using a real-time reverse transcriptase-polymerase chain reaction (RT-PCR). Total RNA wasisolated from two 4-μm sections of each paraffin-embedded tumor using an RNeasy FFPE kit (Qiagen,GmbH, Hilden, Germany), and cDNAs were synthesizedaccording to the manufacturer’s instructions (RocheDiagnostics). cDNAs were then amplified using a Light-Cycler and LightCycler FastStart DNA Master SYBRGreen I according to the manufacturer’s instructions(Roche Diagnostics). The primer sequences for mouseERa were 5’-AAAGCTGGCCTGACTCTG-3’ and 5’-GATGCTCCATGCCTTTGT-3’. The primer sequencesfor mouse glyceraldehyde-3-phosphate dehydrogenase(GAPDH), which was used as an internal control, were5’-TGGCCTTCCGTGTTCCTACC-3’ and 5’-AGCC-CAAGATGCCCTTCAGT-3’. The primer sequences ofERa and GAPDH were determined based on data fromthe GenBank in the National Institutes of Health, USA.The product length was 100 bp for ERa and 135 bp forGAPDH. The levels of ERa mRNA were calculatedusing a 2-ΔΔCt method [25]. The method is based on thefact that the difference in threshold cycles (ΔCt)between the gene of interest (ERa) and housekeepinggene GAPDH is proportional to the relative expressionlevel of the gene of interest.

Statistical analysisSignificant differences in the quantitative data betweenthe groups were analyzed using the Student’s t-test viathe method of Welch, which provides for insufficienthomogeneity of variance. The differences in metastaticincidence were examined by Fisher’s exact probabilitytest, with P < 0.05 or P < 0.01 considered to represent astatistically significant difference.

ResultsER expression of mammary carcinoma cellsIn the BJMC3879luc2 mammary carcinoma cell line, wes-tern blots showed ERa expression between 50 and 64kDa, whereas the MCF-7 human breast cancer cell lineexpressed the 66-kDa form (Figure 1A). ERb was notdetected by western blots (data not shown). Immuno-fluorescence staining demonstrated that this smallerform of ERa was localized to the cytoplasm (Figure 1B).Cell proliferation was significantly increased by the addi-tion of 10 nM of E2. However, other E2 concentrationsdid not change cell proliferation, with the exception ofthe highest concentration (100 μM), which was cytotoxic

(Figure 1C). When tumor cells were implanted intofemale mice and male mice, the resulting tumor volumein the female mice was slightly larger, but the differencewas not statistically significant (Figure 1D). However, thelevels of ERa mRNA in the mammary tumors were sig-nificantly elevated in the females as compared to themales (Figure 1E).

In vitro raloxifene studyCell viabilityCell viability of BJMC3879luc2 mammary cancer cellswas significantly decreased after 48 h of treatment withmore than 10 μM raloxifene (Figure 2A). The concen-tration of raloxifene in the in vitro study was deter-mined to be 20 μM based on cell growth in the IC50concentration. BJMC3879luc2 cells treated with 20 μMraloxifene for 48 h showed a greater number of apopto-tic cells by TUNEL staining as compared to control(data not shown).Caspase activitiesSignificantly elevated activities of caspase-3, caspase-8and caspase-9 were observed in BJMC3879luc2 cellstreated with raloxifene for 24 h (Figure 2B) and 48 h(data not shown), as compared to the respective con-trols. However, the activities of caspase-12 did not showsignificant differences between control cells and raloxi-fene-treated cells (Figure 2B).Release of cytochrome cCytochrome c protein levels in cytosolic fractions weresignificantly elevated in cells treated with raloxifene for48 h (Figure 2C). These findings strongly suggest theengagement of the mitochondria-mediated apoptoticpathway.Bid cleavageSince caspase-8 activities were elevated, we examinedwhether caspase-8-Bid cleavage via the mitochondrialpathway occurred by performing western blots for Bid.Full-length Bid (22 kDa) was detected in control cellsand in cells treated with raloxifene for 48 h (Figure 2D).No cleaved Bid was found.Caspase inhibitor experimentTo determine whether caspase activation is necessary toinduce raloxifene-induced apoptosis, a caspase inhibitorexperiment was conducted. The recovery of cell viabilityoccurred in cells treated with all caspase inhibitors andraloxifene as compared with raloxifene alone for 48 h(Figure 2E).Cell cycle and DNA synthesisAs measured by flow cytometry, 48 h exposure to 20 μMraloxifene induced a significant elevation in the numbersof cells in the G1-phase as compared with control cells(Figure 2F). There was also a significant reduction in theS-phase population in raloxifene-treated cell suspensions(Figure 2F). DNA synthesis in BJMC3879luc2 cells treated

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 5 of 14

with raloxifene for 48 h, as assessed by BrdU incorpora-tion, was significantly decreased (Figure 2G).

In vivo raloxifene studyBody weights and mammary tumor growthBody weight changes in control and raloxifene-treatedmice bearing mammary tumors are shown in Figure3A. The weights of mice treated with raloxifene (18mg/kg or 27 mg/kg) were significantly lower thanthose of control mice throughout of the experiment.At the end of the study, the weight differences betweenthe control group and the raloxifene-treated animalswere 8~10%. One mouse from each group died atweek 6 due to the mammary cancer metastasis. Onemouse from the 27 mg/kg group died accidentally

from an overdose of anesthesia when the osmoticmini-pumps were changed.Tumor volumes are presented in Figure 3B. Tumor

growth, as inferred by computed volume, was significantlyinhibited in the 18 and 27 mg/kg groups from week 3 tothe end of the experiment when compared with controls.By the end of the experiment, the average tumor volumein control animals was 1500 ± 183 mm3, while the averagetumor volume of mice that received raloxifene was 729 ±277 mm3 (18 mg/kg) and 654 ± 161 mm3 (27 mg/kg).Metastasis of mammary carcinomasBioluminescence imaging showed a tendency for meta-static expansion to be decreased in mice treated withraloxifene (Figure 4B, C) as compared to control animals(Figure 4A).

Figure 1 ER expression of mammary carcinoma cells. A. Western blots of ERa in mammary carcinoma BJMC3879luc2 cells showed bandsbetween 50 and 64 kDa, but the MCF-7 human breast cancer cell line showed a ~66 kDa form. ERb was not detected by western blots. B. Inimmunofluorescence staining, this smaller form showed cytoplasmic localization (green). Nuclear stain was conducted with PI (red). Bar = 10 μm.C. Cell proliferation was significantly increased by the addition of 10 nM E2 (**P < 0.01), but cell proliferation was not changed in any otherconcentration of E2, with the exception of the highest concentration of 100 μM, which is toxic. In the case of BJMC3879luc2-implanted tumorsin mice, the tumor volume was slightly bigger in the female mice as compared to the male mice (D), and ERa mRNA levels were significantlyhigher in the implanted tumors in females than those in males (*P < 0.05) (E).

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 6 of 14

Figure 2 In vitro raloxifene study. A. Cell viability was significantly decreased in mouse mammary carcinoma BJMC3879luc2 cells (1 × 104 cellsand 5 × 104 cells/well) treated with more than 10 μM of raloxifene for 48 h (**P < 0.01). The IC50 concentration was determined to be 20 μM;therefore, 20 μM raloxifene for 48 h-incubation was used for in vitro studies. Ten samples from each dosage of raloxifene were examined. B.Caspase activities were evaluated according to luminescent assay. Activities of caspase-3, caspase-8 and caspase-9 (but not caspase-12) weresignificantly elevated in BJMC3879luc2 cells treated with 20 μM raloxifene for 48 h (**P < 0.01). Three samples each of control and raloxifene-treated cells were examined. C. Cytochrome c in the cytosolic fraction, as determined by ELISA, was significantly increased in cells treated withraloxifene for 48 h as compared to the control levels (*P < 0.05). Six samples from control cells and five samples from raloxifene-treated cellswere examined. D. Western blots of Bid (22 kDa) in control cells and cells treated with raloxifene for 48 h were similar (upper panel). Cleaved Bidwas not observed after raloxifene treatment. b-Actin served as an internal control (lower panel). E. In BJMC3879luc2 cells treated with raloxifenefor 48 h, cell viabilities were significantly increased by the broad-spectrum caspase inhibitor z-VAD-fmk, the caspase-3 specific inhibitor Ac-DNLD-CHO, the caspase-8 specific inhibitor z-IETD-fmk, and the caspase-9 specific inhibitor z-LETD-fmk at 10 or 100 μM (*P < 0.05; **P < 0.01). Sixsamples each of control and raloxifene-treated cells were examined. F. Cell-cycle analysis showed that raloxifene induced arrest in the G1-phaseand inhibition of the S-phase in metastatic mouse mammary carcinoma BJMC3879luc2 cells (**P < 0.01). Three samples each of control andraloxifene-treated cells were examined. G. Levels of DNA synthesis, as assessed by BrdU incorporation rates, were significantly decreased in thecells treated with raloxifene for 48 h (**P < 0.01). rlu: relative luminescent unit. Data presented are means ± SD values. Four samples each ofcontrol and raloxifene-treated cells were examined.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 7 of 14

Histopathologically, the mammary carcinomas inducedby BJMC3879luc2 cell inoculation proved to be moder-ately differentiated adenocarcinomas (Figure 4D) thatcontained mutated p53 as inferred by immunohisto-chemistry (Figure 4E).

Lymph node metastasisRepresentative lymph node metastases are shown in Fig-ures 4F and 4G. Lymph node metastasis occurred in allmice independent of groups. However, the number ofmetastasis-positive lymph nodes per mouse was signifi-cantly decreased in the 27 mg/kg group as compared tothe control group (Figure 3C).

Lung metastasisLung metastasis occurred in all mice. The number oflung metastatic foci (>200 μm) per mouse tended todecreases in the raloxifene-treated groups, although thedecrease was not statistically significant (Figure 3D).However, the metastatic foci tended to be smaller in theraloxifene-treated groups (Figure 4I) than in the controlanimals (Figure 4H). In addition, there was no differencein the uterine endometrium of the control mice and theraloxifene-treated mice (Figure 4J, K).Apoptosis and DNA synthesis in mammary cancersResults of the quantitative analysis for apoptosis inlesions, as assessed by the TUNEL assay, are shown inFigure 5A. The number of TUNEL-positive cells wassignificantly increased in tumors from the 18 and 27mg/kg groups (Figure 6B) as compared to the tumorsfrom control mice (Figure 6A). Immunohistochemistrydemonstrated that the expression of the active forms ofcaspase-3 and caspase-9 were much higher in mammarytumors treated with raloxifene (Figure 6D, F) than inthe untreated control tumors (Figure 6C, E), suggestingthat mitochondria-mediated apoptosis occurred in mam-mary tumor tissues exposed to raloxifene in vivo, too.DNA synthesis levels in mammary carcinomas of

raloxifene-treated mice (18 and 27 mg/kg), as inferredby BrdU labeling indices, are shown in Figure 5B. Levelsof DNA synthesis in tumors were significantly decreasedin the 18 and 27 mg/kg groups (Figure 5B and 6G, H).Blood microvascular density and lymphatic vessels inmammary cancersMicrovessel density, as determined by immunohisto-chemical analysis with the blood vessel endothelial cellmarker CD31, showed no statistically significant differ-ence between control mice and raloxifene-treated mice(Figure 5C).The lymphatic vessels in mammary tumors were

stained with anti-podoplanin antibody, as demonstratedin Figures 6I and 6J. There were tumor cells within thelumina of dilated lymphatic vessels of tumors in bothcontrol (Figure 6I) and raloxifene-treated animals

Figure 3 Body weights, tumor volumes and multiplicity ofmetastasis in mammary carcinomas. Raloxifene was administeredwith mini-osmotic pumps. Each group consisted of 10 mice. A.Body weights of mice treated with 18 and 27 mg/kg/day raloxifenewere significantly decreased throughout the experiment ascompared with the control group, but only by 10% (*P < 0.05; **P <0.01). B. Tumor volumes in the 18 and 27 mg/kg/day groups beganto decrease significantly as compared to the control values startingat week 3, and the differences became even more pronounced bythe termination of the experiment (week 6) (*P < 0.05; **P < 0.01).C. Multiplicity of lymph node metastasis was significantly decreasedin the 27 mg/kg raloxifene group (*P < 0.05). D. Multiplicity of lungmetastasis tended to be reduced in the 27 mg/kg raloxifene group.Data are presented as means ± SD.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 8 of 14

Figure 4 Bioluminescent imaging and histopathological findings. Bioluminescent imaging in five representative mice from each group (A,control; B, 18 mg/kg; C, 27 mg/kg). Bioluminescent imaging showed a tendency for decreases in the extension of metastasis in raloxifene-treated groups as compared to the control group. D. The implanted mammary carcinomas proved to be moderately differentiatedadenocarcinoma. ×200. E. p53 immunohistochemistry of mammary carcinoma induced by BJMC3879 cell inoculation. Note the nuclear stainingfor abnormal p53 protein, indicating that these cells carry mutant p53. × 400. F. Metastasis to lymph node in control mice (×40, inset). Metastaticcarcinoma cells were filled with sinusoidal space (× 400). G. A lymph node from a mouse given 27 mg/kg raloxifene (× 40, inset). Metastaticcarcinoma cells were filled with subcapsular sinus and sinusoidal space (× 400). H. Metastatic foci in the lung of a control mouse. Manymetastatic foci and small to large nodules were seen. × 40. I. Metastatic foci in the lungs of mice given 27 mg/kg raloxifene. Metastatic lung fociwere much smaller in the 27 mg/kg group than in the control group. × 40. J and K. Uterine endometrium was histopathologically similarbetween control and raloxifene-treated mice. × 100. D and F-K, H&E stain; B, p53 immunohistochemistry.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 9 of 14

Figure 5 Quantitative analyses of apoptosis, cell proliferation, vascular density and ERa expression in mammary carcinomas. A.Apoptotic cell death, assessed by TUNEL assay, was significantly increased in the 18 and 27 mg/kg raloxifene groups (**P < 0.01). B. DNAsynthesis, inferred by BrdU labeling indices, was significantly decreased in the 18 and 27 mg/kg raloxifene groups (**P < 0.01). C. Microvesseldensity in tumors, inferred by CD31-positive endothelium, was similar between the control group and the raloxifene-treated groups. D. Thenumber of dilated lymphatic vessels containing intraluminal tumor cells was significantly lower in groups receiving 18 and 27 mg/kg raloxifenethan in the control group (*P < 0.05). E. Levels of the truncated ERa mRNA tended to be decreased in the raloxifene-treated groups ascompared to the levels of the control groups, but this difference was not significant as measured by real-time RT-PCR. Data are presented asmeans ± SD.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 10 of 14

(Figure 6J). However, the number of dilated lymphaticvessels containing intraluminal tumor cells (arrows inFigure 6I, J) was significantly reduced in mammarytumors of mice given 18 and 27 mg/kg of raloxifene(Figure 5D), indicating a reduction in the number oftumor cells migrating into the lymphatic vessels oftumor tissues.ERa expression of mammary carcinomaImmunohistochemically, adjacent normal mammaryglands in the mammary carcinomas showed nuclearexpression of ERa in the control mice (Figure 6K),while the ERa expression level of normal mammaryglands in the raloxifene-treated mice tended to bedecreased (Figure 6L). In mammary carcinoma tissues,the scattered expression of ERa was observed in thecytoplasm of both control mice (Figure 6M) and raloxi-fene-treated mice (Figure 6N), and the expression leveltended to decrease in the raloxifene-treated mice. In thequantitative analysis, levels of ERa mRNA in mammarycarcinoma tissues tended to decrease in mice treatedwith raloxifene as compared to the control mice, but thedecrease was not statistically significant (Figure 5E).

DiscussionThe present study showed that raloxifene inhibitedtumor growth and multiplicity of metastasis to lymphnodes in a mouse immunocompetent metastatic mam-mary carcinoma model expressing cytoplasmic ERa. Inaddition, tumor tissues from the raloxifene-treated miceshowed elevation of apoptotic cell death, suppression ofDNA synthesis and inhibition of lymphatic vessels con-taining intraluminal cancer cells.The present in vitro studies showed that the ERa

expressed in the mammary carcinoma BJMC3879luc2cells used in this study was between 50 and 64 kDa, whichis smaller than the 66-kDa size of normal ERa, and itshowed a cytoplasmic location. Cell proliferation ofBJMC3879 cells expressing the smaller molecular weightERa was significantly increased, but only by 7%, whenadded to 10 nM estrogen. When BJMC3879luc2 cells wereimplanted into mice, the ERa mRNA levels in the resul-tant tumors were significantly higher in female mice ascompared to the male mice. Thus, although the ERa inthe present study might be functional but weak, furtherinvestigation is necessary to elucidate this point. Recently,a truncated variant of 36-kDa ERa has been identified[26]. This subtype, which is predominantly localized to thecytoplasm and plasma membrane, responds to estrogenand mediates a nongenomic signaling pathway. Althoughthis 36-kDa variant of ERa is apparently different from thepresent ERa, they share similarities in regard to localiza-tion and estrogen response.The results of STAR [9], MORE [10], CORE [11] and

RUTH [12] clinical trials show that raloxifene does not

Figure 6 Apoptosis, cell proliferation, lymphatic vessels withintraluminal tumor cells and ERa expression in mammarycarcinomas. Whereas some TUNEL-positive cells are seen in thetumor of a control mouse (A), many more TUNEL-positive cells areobserved in the tumor of a mouse given 27 mg/kg raloxifene (B).×200. Expression of active caspase-3 (C and D, x200)and caspase-9(E and F, x400) was more prominent in the tumor of a mouse given27 mg/kg raloxifene (D and F) than in a control mouse (C and E).The number of BrdU-labeled cells tended to be lower in the 27 mg/kg raloxifene group (H) than in the control group (G). ×200.Podoplanin-positive lymphatic vessels of a tumor in a controlmouse were often dilated and filled with tumor cells (arrows, I).Raloxifene-treated groups showed a significant reduction in thenumbers of dilated lymphatic vessels containing intraluminal tumorcells. (arrows, J). ×200. The nuclear ERa expression of adjacentnormal mammary glands in the mammary tumor was muchstronger in the control (K) than in the 27 mg/kg raloxifene group(L). ×400. The scattered expression of cytoplasmic ERa was muchstronger in the control group (M) than in the 27 mg/kg raloxifenegroup (N). ×400. A and B, TUNEL stain; C and D, active caspase-3immunohistochemistry; E and F, active caspase-9immunohistochemistry; G and H, BrdU immunohistochemistry; I and J,podoplanin immunohistochemistry. K-N, ERa immunohistochemistry.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 11 of 14

reduce the risk of ER-negative invasive breast cancer.Therefore, the fact that raloxifene exerted antimetastaticeffects on mammary cancer expressing the cytoplasmicform of ERa may be an important finding with clinicalapplications. The question was raised as to why raloxi-fene exerted antitumor effects on mammary tumors thatexpressed the cytoplasmically located ERa in the presentstudy. ER lacks known functional motifs that wouldallow for nongenomic mechanisms of estrogen action[27]. Raloxifene acts on both nuclear ERa and cytoplas-mic ERa (nongenomic action) [28]. In this case, raloxi-fene does not target the estrogen response element;rather, it targets the raloxifene response element [29]. Itwas previously reported that estrogen activates cell pro-liferation in even ER-negative human breast cancer cellsMDA-MB231 via GPR30, a member of the G protein-coupled receptor superfamily [30]. Thus, raloxifene canact by nongenomic mechanisms independent of ER,indicating the complexity and variety of SERMs. Thebiological effects of raloxifene decrease the ER levels[8,31]. In fact, in the present in vivo study, the mRNAlevels of the truncated ERa in mammary tumors ofraloxifene-treated mice showed a tendency to bedecreased as compared to the levels in control mice. Itis possible that raloxifene acts on the present mammarycancer model. In addition, ERa and b have been pre-viously localized to mitochondria in various tissues[32,33]. In the present study, immunohistochemicallocalization of the truncated ERa revealed scatteredexpression in the cytoplasm, suggesting mitochondriallocalization.The present study demonstrated that raloxifene signif-

icantly induced apoptosis in murine mammary carci-noma cells both in vitro and in vivo. There are twopathways currently proposed to play major roles in reg-ulating apoptosis in mammalian cells: a pathwaymediated by death receptor (extrinsic pathway; execu-tion by caspase-8) and a pathway mediated by mito-chondria (intrinsic pathway; execution by caspase-9)[34]. Caspase-3 is a final executor of apoptosis. Many ofthe apoptosis signals are transduced to the mitochondriaand decrease the mitochondrial membrane potential,which leads to the release of cytochrome c from themitochondrial lumen into the cytoplasm. The releasedcytochrome c binds to the apoptosis protease-activatingfactor-1 (Apaf-1), and this complex activates caspase-9.Caspase-8 also has a cross-talk pathway to the mito-chondria pathway through the cleavage of Bid [34].In the present in vitro study, increases in caspase

activities (caspase-3, -8 and -9) and cytosolic cyto-chrome c levels were found in mammary carcinomacells treated with raloxifene, suggesting that raloxifeneat least induced mitochondria-mediated apoptosis.Indeed, mammary cancer tissues of mice treated with

raloxifene showed strong expression of active caspase-3and -9 (cleaved forms), demonstrating that mitochon-dria-mediated apoptosis also occurred in vivo. All cas-pase inhibitors involving a caspase-8 inhibitorcompletely rescued raloxifene-induced cell death. How-ever, since Bid cleavage was not observed, cross-talkbetween caspase-8 and Bid may not be involved. Thequestion was raised as to why caspase-8 activityincreased. Caspase-8 participates in ERK activation, andthis regulation is attributed to the Death EffectorDomains (DED) of caspase-8 [35]. Furthermore, a directassociation between ERK and a DED-containing frag-ment of caspase-8, and co-transport of an ERK-caspase-8-DED complex to the nucleus during apoptosis hasbeen reported [36]. The caspase-8-ERK pathway mayalso play a role in raloxifene-induced apoptosis. Furtherinvestigation is required to elucidate this point. In addi-tion, caspase-12 mediates the pathway for cell deathinduced by endoplasmic reticulum stress in mice [37].In the present study, since no elevation in caspase-12activity was seen, the raloxifene-induced apoptosis maynot have involved endoplasmic reticulum stress.In animal carcinogenesis models, raloxifene at 20 mg/

kg/day inhibits the tumor growth of 7, 12-dimethylben-zanthracene-induced mammary carcinomas in rats [5].In mice, orally administered raloxifene (1.5 mg/mouse)reduces the tumor growth of mammary and endometrialcancer [23]. On the assumption that mouse bodyweights are 30 g, the dosage of raloxifene is estimatedto be 50 mg/kg/day in mice. In carcinogenicity studiesin mice and rats, raloxifene (8.7~225 mg/kg/day inmice; 10.4~259 mg/kg/day in rats) is not carcinogenic(company data from Eli Lilly Pharmaceuticals, Indiana-polis, IN, USA). Although the clinical dosages of raloxi-fene in trials are 60 mg or 120 mg/day, a much higherdose of 600 mg/day (estimated as 10 mg/kg/day on theassumption that body weight is 60 kg) has also beenused in clinical studies without adverse side effects[38,39]. Therefore, the doses of raloxifene used in thepresent mouse study (18 and 27 mg/kg/day) are notextremely high, and the dosage levels are considered tobe near the clinical dose. However, low doses of raloxi-fene also exert antitumorigenic effects in animal cancermodels [8].Cancer cells metastasize to distal sites via the lympha-

tic system and the vascular system. The lymphatic capil-laries present in tissues and tumors provide entranceinto the lymphatics, allowing cancer cell migration tothe lymph nodes. In the present study, it was demon-strated that the multiplicity of lymph node metastaseswas decreased in raloxifene-treated mice. This phenom-enon was supported by a significant decrease in thenumber of lymphatic vessels with tumor cells in theirlumina in the raloxifene-treated groups. This finding

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 12 of 14

indicates that raloxifene may have an inhibitory effecton migration into lymphatic vessels. In fact, raloxifenehas been reported to inhibit estrogen-induced cellmigration and invasion through a non-nuclear signalingcascade involving G proteins and the RhoA-associatedkinase [40]. It was also reported that raloxifenedecreases levels of cyclooxygenase-2 and inducible nitricoxide synthase in carrageenan-induced inflammation ofrats [41]. This mechanism could possibly be involved inthe antitumorigenic effects of raloxifene.Neovascularization is also a key process in the growth of

solid tumors, and the growth of both primary tumors andmetastases is thus angiogenesis-dependent [42]. However,in the present study, microvessel density in tumors wassimilar between the control and raloxifene-treated groups,indicating that raloxifene may not have anti-angiogenicaction. However, the microvessel density in the 27 mg/kgraloxifene group was slightly increased. Since raloxifeneinduces cell proliferation and up-regulation of telomeraseactivity in human umbilical vein endothelial cells [43], thiseffect might be involved in the present study. However,since raloxifene did not inhibit angiogenesis in tumors inthe present study, lung metastasis may not have beenstrongly inhibited.The present experiments suggest that raloxifene-

induced apoptosis in BJMC3879Luc2 cells having a p53mutation occurs through a p53-independent mechanism.Since 50% of human cancers have p53 mutations [44],the fact that the raloxifene induces a p53-independentapoptotic response in cancer cells having a p53 muta-tion may be highly relevant to inhibiting many humancancers. In the case of non-functional p53 status, p73,the p53 homologue, may play a role in apoptosisinduction.

ConclusionOur results demonstrated that treatment with raloxifenesignificantly suppresses lymph node metastasis in amouse mammary cancer model expressing cytoplasmicERa. The antimetastatic activity of raloxifene may be acrucial finding with clinical applications, and raloxifenemay be useful as an adjuvant therapy and for the che-moprevention of breast cancer development.

AbbreviationsBrdU: 5’-bromo-2’-deoxyuridine; CORE: Continuing Outcomes Relevant toEvista; DMSO: dimethylsulfoxide; E2: 17-b estradiol; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; LSAB: labeled streptavidin-biotin; MORE: Resultsof other clinical trials of raloxifene, such as the Multiple Outcomes ofRaloxifene Evaluation; MMTV: mouse mammary tumor virus; PBS: Phosphate-buffered saline; RT-PCR: reverse transcriptase-polymerase chain reaction;RUTH: Raloxifene Use for The Heart; SERM: selective estrogen receptormodulator; STAR: Study of Tamoxifen and Raloxifene; TUNEL: terminaldeoxynucleotidyl transferase-mediated dUTP-FITC nick end-labeling.

AcknowledgementsThis investigation was supported by a Grant-in-Aid for Private Universitiesfrom the Ministry of Education, Culture, Sports, Science and Technology(MEXT) of Japan (referred to as Shibata’s Project in the Central ResearchLaboratory of Osaka Medical College). We thank Dr. Shingo Kamoshida(Kobe University Graduate School of Health Sciences) forimmunohistochemistry for caspases and Mr. Teruo Ueno (the CentralResearch Laboratory of Osaka Medical College) for assistance with the cell-cycle analysis. We are also grateful to Ms. Mika Yoshida and Yumi Naimitafor their warm-hearted secretarial assistance.

Author details1Department of Anatomy and Cell Biology, Division of Life Sciences, OsakaMedical College, 2-7 Daigaku-machi, Takatsuki, Osaka 569-8686, Japan.2Laboratory Animal Center, Osaka Medical College, 2-7 Daigaku-machi,Takatsuki, Osaka 569-8686, Japan. 3Department of Systems Bioscience forDrug Discovery, Graduate School of Pharmaceutical Sciences, KyotoUniversity, Kyoto, Japan. 4Research Center for Food Safety, University ofTokyo Graduate School of Agricultural and Life Sciences, Tokyo, Japan.5Department of Gynecology, Osaka Medical College, 2-7 Daigaku-machi,Takatsuki, Osaka 569-8686, Japan. 6Department of Bioscience, NationalCardiovascular Center Research Institute, Suita, Osaka, Japan.

Authors’ contributionsMAS performed the cell culture, animal experiments, Western blots,histopathology and statistical analysis. All in vitro studies (except for Westernblots, cell-cycle analysis and ERa immunofluorescence) were performed byES. Transplantation was performed by JM. Cell-cycle analysis was performedby HK. ERa immunofluorescence was conducted by KA. Experiments wereperformed by MAS, and ZL, MK, MO and YO participated in the design ofthe study. MAS wrote the manuscript. All authors have read and approvedthe final manuscript to be submitted.

Competing interestsThe authors declare that they have no competing interests.

Received: 26 July 2010 Accepted: 19 October 2010Published: 19 October 2010

References1. Fisher B, Costantino JP, Wickerham DL, Redmond CK, Kavanah M,

Cronin WM, Vogel V, Robidoux A, Dimitrov N, Atkins J, et al: Tamoxifen forprevention of breast cancer: report of the National Surgical AdjuvantBreast and Bowel Project P-1 Study. J Natl Cancer Inst 1998,90(18):1371-1388.

2. Rutqvist LE, Johansson H, Signomklao T, Johansson U, Fornander T,Wilking N: Adjuvant tamoxifen therapy for early stage breast cancer andsecond primary malignancies. Stockholm Breast Cancer Study Group. JNatl Cancer Inst 1995, 87(9):645-651.

3. Cuzick J, Powles T, Veronesi U, Forbes J, Edwards R, Ashley S, Boyle P:Overview of the main outcomes in breast-cancer prevention trials.Lancet 2003, 361(9354):296-300.

4. Shang Y, Brown M: Molecular determinants for the tissue specificity ofSERMs. Science 2002, 295(5564):2465-2468.

5. Clemens JA, Bennett DR, Black LJ, Jones CD: Effects of a new antiestrogen,keoxifene (LY156758), on growth of carcinogen induced mammarytumors and on LH and prolactin levels. Life Sci 1983, 32(25):2869-2875.

6. Gottardis MM, Jordan VC: Antitumor actions of keoxifene and tamoxifenin the N-nitrosomethylurea-induced rat mammary carcinoma model.Cancer Res 1987, 47(15):4020-4024.

7. Anzano MA, Peer CW, Smith JM, Mullen LT, Shrader MW, Logsdon DL,Driver CL, Brown CC, Roberts AB, Sporn MB: Chemoprevention ofmammary carcinogenesis in the rat: combined use of raloxifene and 9-cis-retinoic acid. J Natl Cancer Inst 1996, 88(2):123-125.

8. Janakiram NB, Steele VE, Rao CV: Estrogen receptor-beta as a potentialtarget for colon cancer prevention: chemoprevention of azoxymethane-induced colon carcinogenesis by raloxifene in F344 rats. Cancer Prev Res(Phila Pa) 2009, 2(1):52-59.

9. Vogel VG: The NSABP Study of Tamoxifen and Raloxifene (STAR) trial.Expert Rev Anticancer Ther 2009, 9(1):51-60.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 13 of 14

10. Cummings SR, Eckert S, Krueger KA, Grady D, Powles TJ, Cauley JA,Norton L, Nickelsen T, Bjarnason NH, Morrow M, et al: The effect ofraloxifene on risk of breast cancer in postmenopausal women: resultsfrom the MORE randomized trial. Multiple Outcomes of RaloxifeneEvaluation. Jama 1999, 281(23):2189-2197.

11. Martino S, Cauley JA, Barrett-Connor E, Powles TJ, Mershon J, Disch D,Secrest RJ, Cummings SR: Continuing outcomes relevant to Evista: breastcancer incidence in postmenopausal osteoporotic women in arandomized trial of raloxifene. J Natl Cancer Inst 2004, 96(23):1751-1761.

12. Barrett-Connor E, Mosca L, Collins P, Geiger MJ, Grady D, Kornitzer M,McNabb MA, Wenger NK: Effects of raloxifene on cardiovascular eventsand breast cancer in postmenopausal women. N Engl J Med 2006,355(2):125-137.

13. Lee WL, Cheng MH, Chao HT, Wang PH: The role of selective estrogenreceptor modulators on breast cancer: from tamoxifen to raloxifene.Taiwan J Obstet Gynecol 2008, 47(1):24-31.

14. Kuroishi T, Tominaga S: Epidemiol. Breast Cancer. Jpn J Cancer Chemother2001, 28:168-173.

15. Shibata MA, Morimoto J, Otsuki Y: Suppression of murine mammarycarcinoma growth and metastasis by HSVtk/GCV gene therapy using invivo electroporation. Cancer Gene Ther 2002, 9:16-27.

16. Shibata MA, Ito Y, Morimoto J, Otsuki Y: Lovastatin inhibits tumor growthand lung metastasis in mouse mammary carcinoma model: a p53-independent mitochondrial-mediated apoptotic mechanism.Carcinogenesis 2004, 25:1887-1898.

17. Shibata MA, Morimoto J, Shibata E, Otsuki Y: Combination therapy withshort interfering RNA vectors against VEGF-C and VEGF-A suppresseslymph node and lung metastasis in a mouse immunocompetentmammary cancer model. Cancer Gene Ther 2008, 15(12):776-786.

18. Shibata MA, Shibata E, Morimoto J, Eid NAS, Tanaka Y, Watanabe M,Otsuki Y: An immunocompetent murine model of metastatic mammarycancer accessible to bioluminescence imaging. Anticancer Res 2009,29:4389-4396.

19. Shibata MA, Liu ML, Knudson MC, Shibata E, Yoshidome K, Bandy T,Korsmeyer SJ, Green JE: Haploid loss of bax leads to acceleratedmammary tumor development in C3(1)/SV40-TAg transgenic mice:reduction in protective apoptotic response at the preneoplastic stage.EMBO J 1999, 18:2692-2701.

20. Yoshimori A, Sakai J, Sunaga S, Kobayashi T, Takahashi S, Okita N,Takasawa R, Tanuma S: Structural and functional definition of thespecificity of a novel caspase-3 inhibitor, Ac-DNLD-CHO. BMC Pharmacol2007, 7:8.

21. Tanuma S, Yoshimori A, Takasawa R: Genomic drug discovery forapoptosis regulation using a new computer screening amino acidcomplement wave method. Biol Pharm Bull 2004, 27(7):968-973.

22. Shibata MA, Akao Y, Shibata E, Nozawa Y, Ito T, Mishima S, Morimoto J,Otsuki Y: Vaticanol C, a novel resveratrol tetramer, reduces lymph nodeand lung metastases of mouse mammary carcinoma carrying p53mutation. Cancer Chemother Pharmacol 2007, 60:681-691.

23. O’Regan RM, Osipo C, Ariazi E, Lee ES, Meeke K, Morris C, Bertucci A,Sarker MA, Grigg R, Jordan VC: Development and therapeutic options forthe treatment of raloxifene-stimulated breast cancer in athymic mice.Clin Cancer Res 2006, 12(7 Pt 1):2255-2263.

24. Gorrin-Rivas MJ, Arii S, Furutani M, Mizumoto M, Mori A, Hanaki K,Maeda M, Furuyama H, Kondo Y, Imamura M: Mouse macrophagemetalloelastase gene transfer into a murine melanoma suppressesprimary tumor growth by halting angiogenesis. Clin Cancer Res 2000,6(5):1647-1654.

25. Livak KJ, Schmittgen TD: Analysis of relative gene expression data usingreal-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods2001, 25:402-408.

26. Lee LM, Cao J, Deng H, Chen P, Gatalica Z, Wang ZY: ER-alpha36, a novelvariant of ER-alpha, is expressed in ER-positive and -negative humanbreast carcinomas. Anticancer Res 2008, 28(1B):479-483.

27. Musgrove EA, Sutherland RL: Biological determinants of endocrineresistance in breast cancer. Nat Rev Cancer 2009, 9(9):631-643.

28. Simoncini T, Genazzani AR, Liao JK: Nongenomic mechanisms ofendothelial nitric oxide synthase activation by the selective estrogenreceptor modulator raloxifene. Circulation 2002, 105(11):1368-1373.

29. Yang NN, Venugopalan M, Hardikar S, Glasebrook A: Identification of anestrogen response element activated by metabolites of 17beta-estradioland raloxifene. Science 1996, 273(5279):1222-1225.

30. Filardo EJ, Quinn JA, Bland KI, Frackelton AR Jr: Estrogen-inducedactivation of Erk-1 and Erk-2 requires the G protein-coupled receptorhomolog, GPR30, and occurs via trans-activation of the epidermalgrowth factor receptor through release of HB-EGF. Mol Endocrinol 2000,14(10):1649-1660.

31. Dowsett M: Preoperative models to evaluate endocrine strategies forbreast cancer. Clin Cancer Res 2003, 9(1 Pt 2):502S-510S.

32. Chi A, Chen X, Chirala M, Younes M: Differential expression of estrogenreceptor beta isoforms in human breast cancer tissue. Anticancer Res2003, 23(1A):211-216.

33. Cammarata PR, Chu S, Moor A, Wang Z, Yang SH, Simpkins JW: Subcellulardistribution of native estrogen receptor alpha and beta subtypes incultured human lens epithelial cells. Exp Eye Res 2004, 78(4):861-871.

34. Hengartner MO: The biochemistry of apoptosis. Nature 2000, 407:770-776.35. Finlay D, Vuori K: Novel noncatalytic role for caspase-8 in promoting SRC-

mediated adhesion and Erk signaling in neuroblastoma cells. Cancer Res2007, 67(24):11704-11711.

36. Yao Z, Duan S, Hou D, Heese K, Wu M: Death effector domain DEDa, aself-cleaved product of caspase-8/Mch5, translocates to the nucleus bybinding to ERK1/2 and upregulates procaspase-8 expression via a p53-dependent mechanism. Embo J 2007, 26(4):1068-1080.

37. Morishima N, Nakanishi K, Takenouchi H, Shibata T, Yasuhiko Y: Anendoplasmic reticulum stress-specific caspase cascade in apoptosis.Cytochrome c-independent activation of caspase-9 by caspase-12. J BiolChem 2002, 277(37):34287-34294.

38. Draper MW, Flowers DE, Huster WJ, Neild JA, Harper KD, Arnaud C: Acontrolled trial of raloxifene (LY139411681) HCl: impact on boneturnover and serum lipid profile in healthy postmenopausal women. JBone Miner Res 1996, 835-842.

39. Boss SM, Huster WJ, Neild JA, Glant MD, Eisenhut CC, Draper MW: Effects ofraloxifene hydrochloride on the endometrium of postmenopausalwomen. Am J Obstet Gynecol 1997, 177(6):1458-1464.

40. Flamini MI, Fu XD, Sanchez AM, Giretti MS, Garibaldi S, Goglia L,Pisaneschi S, Tosi V, Genazzani AR, Simoncini T: Effects of raloxifene onbreast cancer cell migration and invasion through the actincytoskeleton. J Cell Mol Med 2009, 13(8B):2396-2407.

41. Esposito E, Iacono A, Raso GM, Pacilio M, Coppola A, Di Carlo R, Meli R:Raloxifene, a selective estrogen receptor modulator, reducescarrageenan-induced acute inflammation in normal and ovariectomizedrats. Endocrinology 2005, 146(8):3301-3308.

42. Folkman J: Angiogenesis-dependent diseases. Semin Oncol 2001,28(6):536-542.

43. Doshida M, Ohmichi M, Tsutsumi S, Kawagoe J, Takahashi T, Du B, Mori-Abe A, Ohta T, Saitoh-Sekiguchi M, Takahashi K, et al: Raloxifene increasesproliferation and up-regulates telomerase activity in human umbilicalvein endothelial cells. J Biol Chem 2006, 281(34):24270-24278.

44. Greenblatt MS, Bennett WP, Hollstein M, Harris CC: Mutations in the p53tumor suppressor gene: clues to cancer etiology and molecularpathogenesis. Cancer Res 1994, 54:4855-4878.

Pre-publication historyThe pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2407/10/566/prepub

doi:10.1186/1471-2407-10-566Cite this article as: Shibata et al.: Raloxifene inhibits tumor growth andlymph node metastasis in a xenograft model of metastatic mammarycancer. BMC Cancer 2010 10:566.

Shibata et al. BMC Cancer 2010, 10:566http://www.biomedcentral.com/1471-2407/10/566

Page 14 of 14